Tweet
Hi Nathan Deans:
You will have to bandlimit your signal before you can sample it (true for any
signal to avoid high frequency noise). You need to find out what the highest
frequency in the EMG signal that you are interested in is (can be higher for
patients with certain neuro and/or muscular conditions). Then you need to be
sure that you are sampling well above the nyquist sampling criteria of twice the
maximum frequency. For eg. if your bandlimit your EMG signal to 1000 Hz. Twice
the maximum frequency is 2000 Hz. Sampling at 3000 Hz will give a good signal
on which you can perform FFT.
Hope this helps
Vineet Gupta, Ph.D.
Vineet Gupta, Ph.D.
Senior Research Engineer
Assistive Technology Research Center/Rehabilitation Engineering
National Rehabilitation Hospital
102, Irving Street, NW
Washington, DC 20010
Phone: (202) 877-1554 (Office)
(301) 847-6716 (Home)
Email: vxg4@mhg.edu (Office)
seema@sprintmail.com (Home)
Nathan Deans on 03/20/2000 02:11:39 AM
Please respond to Nathan Deans
To: BIOMCH-L%NIC.SURFNET.NL@internet.mhg.edu
cc: (bcc: Vineet Gupta/WHC/MedStar)
Subject: Fourrier Analysis of EMG
Hi Everybody,
I have a question concerning my sampling frequency for EMG data collection
and fast fourrier transformation. The current information that has been
provided to me suggests that i need a sampling frequence of 5000Hz in order
to perform an FFT analysis of the EMG signal.
Unfortunately the equipment performing the data collection is not capable of
handling the large influx of information (4 channels, 5sec of data - 100 000
data points).
The question therefore is how far can i reduce my sampling frequency before
i begin corrupting the integrity of the FFT capabilities? I have played
around with the sampling frequency and i believe i can get what i need at
3000H?
Any help or suggestions would be greatly appreciated.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Nathan Deans (BHMS)(Ex. Sci.)(Hons)
School of Exercise Science & Sport Management
Southern Cross University
P.O.Box 157
Lismore, NSW, 2480, Australia
+61 2 6620 3231 Office
+61 2 6620 3880 Fax
ndeans@scu.edu.au
http://sessm.scu.edu.au/nathandeans (updated)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
---------------------------------------------------------------
To unsubscribe send SIGNOFF BIOMCH-L to LISTSERV@nic.surfnet.nl
For information and archives: http://isb.ri.ccf.org/biomch-l
---------------------------------------------------------------
---------------------------------------------------------------
To unsubscribe send SIGNOFF BIOMCH-L to LISTSERV@nic.surfnet.nl
For information and archives: http://isb.ri.ccf.org/biomch-l
---------------------------------------------------------------
You will have to bandlimit your signal before you can sample it (true for any
signal to avoid high frequency noise). You need to find out what the highest
frequency in the EMG signal that you are interested in is (can be higher for
patients with certain neuro and/or muscular conditions). Then you need to be
sure that you are sampling well above the nyquist sampling criteria of twice the
maximum frequency. For eg. if your bandlimit your EMG signal to 1000 Hz. Twice
the maximum frequency is 2000 Hz. Sampling at 3000 Hz will give a good signal
on which you can perform FFT.
Hope this helps
Vineet Gupta, Ph.D.
Vineet Gupta, Ph.D.
Senior Research Engineer
Assistive Technology Research Center/Rehabilitation Engineering
National Rehabilitation Hospital
102, Irving Street, NW
Washington, DC 20010
Phone: (202) 877-1554 (Office)
(301) 847-6716 (Home)
Email: vxg4@mhg.edu (Office)
seema@sprintmail.com (Home)
Nathan Deans on 03/20/2000 02:11:39 AM
Please respond to Nathan Deans
To: BIOMCH-L%NIC.SURFNET.NL@internet.mhg.edu
cc: (bcc: Vineet Gupta/WHC/MedStar)
Subject: Fourrier Analysis of EMG
Hi Everybody,
I have a question concerning my sampling frequency for EMG data collection
and fast fourrier transformation. The current information that has been
provided to me suggests that i need a sampling frequence of 5000Hz in order
to perform an FFT analysis of the EMG signal.
Unfortunately the equipment performing the data collection is not capable of
handling the large influx of information (4 channels, 5sec of data - 100 000
data points).
The question therefore is how far can i reduce my sampling frequency before
i begin corrupting the integrity of the FFT capabilities? I have played
around with the sampling frequency and i believe i can get what i need at
3000H?
Any help or suggestions would be greatly appreciated.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Nathan Deans (BHMS)(Ex. Sci.)(Hons)
School of Exercise Science & Sport Management
Southern Cross University
P.O.Box 157
Lismore, NSW, 2480, Australia
+61 2 6620 3231 Office
+61 2 6620 3880 Fax
ndeans@scu.edu.au
http://sessm.scu.edu.au/nathandeans (updated)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
---------------------------------------------------------------
To unsubscribe send SIGNOFF BIOMCH-L to LISTSERV@nic.surfnet.nl
For information and archives: http://isb.ri.ccf.org/biomch-l
---------------------------------------------------------------
---------------------------------------------------------------
To unsubscribe send SIGNOFF BIOMCH-L to LISTSERV@nic.surfnet.nl
For information and archives: http://isb.ri.ccf.org/biomch-l
---------------------------------------------------------------